RNAs in the cell are typically found in RNA:protein complexes, but there are few comprehensive descriptions of how the proteins recognize their RNA targets. The U1A/U2B3 family of RNA binding proteins is present in the splicing snRNPs of all eukaryotes, but has evolved via at least three paths from a common ancestor. In higher eukaryotes, two proteins are found in the U1 and U2 snRNPs: U1A binds to U1 snRNA stemloop II and U2B3 binds to U2 snRNA stemloop IV. In insects, there is only one protein that binds both RNAs, which in Drosophila sp is SNF. In lower eukaryotes, such as C elegans, there are again two proteins, but each protein can bind to each stemloop. These three branches of the family tree are equidistant from the common ancestor, and so represent three different solutions to the problem of specific RNA recognition. Equally important to the evolutionary adaptation, however, are the RNA stemloops. Although they all share a common six nucleotide sequence, that sequence is embedded in different contexts within the larger stemloop, such that SLII/SLIV of worms is very unlikely to be recognized by human proteins. This example of co- evolution of RNA and protein offers an unprecedented opportunity to study the molecular details of adaptation and to characterize a highly conserved example of RNA binding proteins. Each protein:RNA complex is investigated in a Specific Aim. Aim 1 is devoted to Drosophila SNF solution structure and dynamics determined by NMR, RNA binding, and interactions of SNF with the Drosophila U2A2 auxiliary protein. Aim 2 does the same for human U2B3, for although there is a cocrystal of U2B3/SLIV/U2A2, there are no biochemical data describing its solution properties or RNA binding ability. In Aim 3, the stemloops II and IV from human, fly, and worm are studied in order to understand their solution structure and dynamics. These are uncharacteristically large loops that must be flexible to drape over the protein. This work uses absorbance, fluorescence, NMR, and computational methods. Aim 4 encompasses C elegans U1A and U2B3; their structures will be determined by either NMR or crystallography, their backbone dynamics by NMR, and RNA binding by fluorescence and biochemical assays. All proteins will be studied computationally to observe their rapid dynamics and their slower motions. With this compendium of information, a global description of co-evolution will emerge. A first hypothesis is that the body of the proteins are conserved in structure and sequence (they are all RNA recognition motifs, RRMs), but their loops (Loop 3 in particular) contain interspersed unique amino acids that both recognize and discriminate among RNAs. The dynamics and conformational sampling of Loop 3 are also controlled by sequence, and those properties are key to RNA binding. Swapping loops should alter RNA recognition. This is a necessarily simplified testable hypothesis that will be fine-tuned as data are acquired.